WO2021149509A1 - 撮像装置、撮像方法、及び、プログラム - Google Patents
撮像装置、撮像方法、及び、プログラム Download PDFInfo
- Publication number
- WO2021149509A1 WO2021149509A1 PCT/JP2021/000445 JP2021000445W WO2021149509A1 WO 2021149509 A1 WO2021149509 A1 WO 2021149509A1 JP 2021000445 W JP2021000445 W JP 2021000445W WO 2021149509 A1 WO2021149509 A1 WO 2021149509A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- imaging
- virtual
- subject
- image
- imaging position
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/60—Control of cameras or camera modules
- H04N23/698—Control of cameras or camera modules for achieving an enlarged field of view, e.g. panoramic image capture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/10—Geometric effects
- G06T15/20—Perspective computation
- G06T15/205—Image-based rendering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
- G06T17/05—Geographic models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/24—Indexing scheme for image data processing or generation, in general involving graphical user interfaces [GUIs]
Definitions
- the present technology relates to an image pickup device, an image pickup method, and a program, and more particularly to an image pickup device, an image pickup method, and a program that enable an image taken from a desired position to be easily obtained.
- Patent Document 1 as a technique for obtaining a virtual image captured from a virtual virtual imaging position different from the actual imaging position, a subject is imaged from various imaging positions using a large number of imaging devices, and the subject is imaged. A technique for generating highly accurate three-dimensional data from an captured image obtained by imaging is described.
- Patent Document 1 In the technique described in Patent Document 1, it is necessary to arrange a large number of imaging devices at various positions. Therefore, in many cases, it cannot be easily realized due to the cost of the imaging device and the labor required for installation.
- This technology was made in view of such a situation, and makes it possible to easily obtain an image captured from a desired position.
- the imaging device or program of the present technology uses the distance information from the imaging position to the subject and the model information to capture the subject from the imaging position, and from a virtual imaging position different from the imaging position.
- An imaging device including a generation unit that generates a virtual image of a subject, or a program for operating a computer as such an imaging device.
- the subject is imaged from a virtual imaging position different from the imaging position from the captured image obtained by capturing the subject from the imaging position using the distance information from the imaging position to the subject and the model information. It is an imaging method including generating a virtual image.
- the captured image obtained by capturing the subject from the imaging position is a virtual image different from the imaging position.
- a virtual image of the subject is generated from the imaging position.
- the imaging device may be an independent device or an internal block constituting one device.
- the program can be provided by transmitting via a transmission medium or by recording on a recording medium.
- FIG. 1 is a diagram showing an example of an imaging situation by an imaging device.
- FIG. 1 the imaging situation is shown by the third trigonometry.
- FIG. 1 a person stands in front of the building when viewed from the image pickup device side, and the person and the building are imaged by the image pickup device from the front side of the person.
- FIG. 2 is a diagram showing an example of an image pickup state by the image pickup device and an image captured image in the image pickup state.
- a in FIG. 2 is a top view showing an imaging situation
- B in FIG. 2 shows an captured image captured in the imaging situation of A in FIG.
- the alternate long and short dash line represents the angle of view of the imaging device, and the space within the angle of view is imaged by the imaging device.
- the broken line represents the angle of view occupied by a person who is the main subject (main subject).
- the distance between the person and the building is relatively long compared to the distance between the imaging device and the person (imaging distance) when the imaging device performs imaging. ing. Therefore, in reality, the width of the building behind the person is wider than the width of the person, but in the captured image, the width of the building is narrower than the width of the person. This is because distant objects look small due to the so-called perspective method.
- FIG. 3 is a diagram showing another example of the imaging situation by the imaging device and the captured image captured in the imaging situation.
- a in FIG. 3 is a top view showing an imaging situation
- B in FIG. 3 shows an captured image captured in the imaging situation of A in FIG.
- the alternate long and short dash line represents the angle of view of the imaging device, and the broken line represents the angle of view occupied by the person.
- a telephoto lens with a narrow angle of view (or a zoom lens with a long focal length) is used from an imaging position farther from the subject than in the case of A in FIG.
- a person and a building are imaged as the same subject as the above.
- the content (composition) of the captured image obtained differs depending on the imaging distance between the subject and the imaging device.
- FIG. 4 is a diagram showing another example of the imaging situation by the imaging device.
- FIG. 4 as in FIG. 1, the imaging situation is shown by the third trigonometry.
- the optical axis direction of the imaging device substantially coincides with the direction from the person to the building.
- the captured image captured by the imaging device is an image in which it is difficult to express the sense of distance between the person and the building, as shown in B of FIG. 2 and B of FIG.
- FIG. 4 a person and a building are imaged from above in front of the person with the optical axis of the image pickup device directed toward the person.
- the direction of the optical axis of the imaging device is different from the direction from the person to the building, and it is possible to obtain an captured image expressing a sense of distance between the person and the building.
- FIG. 5 is a diagram showing an captured image obtained by imaging a person and a building from above in front of the person.
- the bird's-eye view image obtained can be obtained as a captured image.
- FIG. 6 is a top view showing an example of an imaging situation in which imaging cannot be performed from a position far from a person.
- FIG. 6 there is a wall on the front side of the person. Therefore, when the person is imaged from the front, the image pickup device cannot be physically moved to the rear side of the wall surface on the front side of the person, so that the person cannot be imaged from a long distance.
- a tripod or stepladder when a person and a building are imaged from above in front of the person, if a tripod or stepladder is used, the image can be taken from above to some extent. However, when using a tripod or stepladder, imaging from a few meters above is the limit. Furthermore, the use of a tripod or stepladder reduces maneuverability at the imaging site.
- the operation of the drone is not always easy, and it is affected by the weather such as rain and wind outdoors.
- drones cannot be used in places where drone flight is restricted or where drone flight is prohibited due to crowded people.
- Patent Document 1 a subject is imaged from various imaging positions using a large number of imaging devices, and arbitrary virtual virtual imaging is performed from three-dimensional data generated using the captured images obtained by the imaging.
- a technique for generating a virtual image (as if) of an image of a subject from a position is described.
- Patent Document 1 it is necessary to arrange a large number of imaging devices at various positions, and the imaging situation as described in Patent Document 1 can be easily performed due to the cost of the imaging devices and the labor required for installation. Often cannot be realized.
- a virtual image obtained by capturing the subject from a virtual imaging position different from the imaging position is generated from the captured image obtained by capturing the subject from the imaging position. This makes it possible to easily obtain a virtual image of a subject taken from a desired virtual imaging position without installing a large number of imaging devices in the present technology.
- a method of generating a virtual image of a subject from a desired virtual imaging position from an image of a subject captured from a certain imaging position is, for example, a wide-angle lens (or focal distance) as in the imaging situation of FIG. 2A. From the image taken from the imaged position at a short distance from the subject using a zoom lens in a shortened state), from the imaged position at a long distance from the subject using a telephoto lens, as shown in the imaging situation of A in FIG. This is a method of generating an captured image as a virtual image.
- FIG. 7 is a diagram illustrating a perspective projection conversion when an image is taken by an imaging device.
- FIG. 7 shows the relationship between the actual subject on the object surface on which the subject exists and the image on the image pickup surface of the image pickup device that performs photoelectric conversion of the image pickup device.
- FIG. 7 is a top view of a subject standing vertically on the ground surface as viewed from above, and the horizontal direction represents a horizontal position (horizontal direction) horizontal to the ground surface.
- the distance from the object surface to the lens of the imaging device (imaging distance between the subject on the object surface and the imaging device) is called the object distance and is expressed by L obj.
- the distance from the lens to the imaging surface is called the image distance and is represented by L img.
- the position on the object surface, that is, the distance from the optical axis of the image pickup device on the object surface is represented by X obj.
- the position on the imaging surface, that is, the distance from the optical axis of the imaging device on the imaging surface is represented by X img.
- the position X img on the imaging surface corresponding to the position X obj of the subject on the object surface can be expressed by the equation (2).
- Equation (2) represents a so-called perspective projection transformation.
- the perspective projection conversion of the equation (2) is performed physically (optically) at the time of actual imaging of the subject by the imaging device.
- the position X obj of the subject on the object surface corresponding to the position X img on the imaging surface can be expressed by the equation (3).
- Equation (3) represents the inverse transformation of the perspective projection transformation (perspective projection inverse transformation).
- the object distance L obj In order to perform the inverse transformation of the perspective projection of the equation (3), the object distance L obj , the image distance L img , and the position X img of the subject on the imaging surface are required.
- the imaging device that captures the subject can recognize (acquire) the image distance L img and the position X img of the subject on the imaging surface.
- an object distance L obj with a resolution of pixel unit or close to that is required.
- any method can be adopted as a method for obtaining the object distance L obj.
- a so-called stereo method can be adopted in which the distance to the subject is calculated from the parallax obtained by using a plurality of image pickup elements that perform photoelectric conversion.
- a method of irradiating the subject with a determined optical pattern and calculating the distance to the subject from the shape of the optical pattern projected on the subject can be adopted.
- a method called ToF (Time of Flight) can be adopted, which calculates the distance to the subject from the time when the reflected light from the subject returns after irradiating the laser beam.
- a method of calculating the distance to the subject can be adopted by using the image plane phase difference method, which is one of the so-called autofocus methods.
- the distance to the subject can be calculated by combining a plurality of the above methods.
- the subject is moved from the virtual imaging position that is different from the actual object distance L obj from the subject by the perspective projection transformation and the perspective projection inverse transformation.
- a method of generating an imaged (if any) virtual image is described.
- FIG. 8 is a diagram showing an example of an imaging situation in which a subject existing on a single object surface is imaged.
- FIG. 8 as in FIG. 1, the imaging situation is shown by the third trigonometry.
- the subject exists on a single object surface, and the object surface is parallel to the imaging surface of the imaging device. Therefore, the object surface is orthogonal to the optical axis of the image pickup apparatus.
- FIG. 9 is a top view showing a state of wide-angle imaging in which a subject is imaged from an imaging position close to the subject using a wide-angle lens in the imaging situation of FIG.
- a subject whose position on the object surface is X obj is imaged from an imaging position separated by an object distance of L obj_W.
- the image distance at the time of wide-angle imaging is L img_W
- the position of the subject on the imaging surface is X img_W .
- FIG. 10 is a top view showing a state of telephoto imaging in which a subject is imaged from an imaging position far from the subject using a telephoto lens in the imaging situation of FIG.
- the position on the object surface is X obj
- the same subject as in the case of the wide-angle imaging of FIG. 9 is imaged from the imaging position separated by the object distance L obj_T.
- the image distance at the time of telephoto imaging is L img_T
- the position of the subject on the imaging surface is X img_T .
- the perspective projection inverse transformation equation (4) can be obtained.
- the perspective projection conversion equation (5) By applying the perspective projection conversion equation (2) to the telephoto imaging of FIG. 10, the perspective projection conversion equation (5) can be obtained.
- X img_T L img_T / L obj_T ⁇ X obj ... (5)
- Equation (6) can be obtained by substituting X obj on the left side of equation (4) into X obj on the right side of equation (5).
- X img_T (L img_T / L obj_T ) ⁇ (L obj_W / L img_W ) ⁇ X img_W ... (6)
- the coefficient k is defined by the equation (7).
- equation (6) can be a simple proportional equation of equation (8).
- wide-angle imaging using a wide-angle lens here, telephoto using a telephoto lens from the position X img_W on the imaging surface in short-range imaging from a short distance.
- Imaging here the position X img_T on the imaging surface in long-distance imaging from a long distance can be obtained.
- the image is taken by long-distance imaging based on the information such as the captured image obtained by the actual imaging by short-distance imaging, the virtual image that will be obtained by the imaging by the long-distance imaging. Information can be obtained.
- imaging from imaging positions with different distances from the subject has been described by taking short-range imaging using a wide-angle lens and long-distance imaging using a telephoto lens as examples. Can be applied when imaging at an arbitrary distance from the subject.
- a lens having another focal length is used based on information such as an image obtained by imaging from a certain imaging position using a lens having a certain focal length. It is possible to obtain information on a captured image (virtual image) obtained when imaging is performed from another imaging position (virtual imaging position) used.
- imaging from a certain imaging position using a lens with a certain focal length is an actual imaging, so it is also called an actual imaging.
- imaging from another imaging position (virtual imaging position) using a lens with another focal length is also called virtual imaging because it is not an actual imaging.
- FIG. 11 is a diagram illustrating an example of a process of obtaining a virtual image obtained by virtual imaging based on information obtained by actual imaging using equation (8).
- the position of the subject on the imaging surface X img_W is the position of the point on the subject in the three-dimensional space that is perspectively projected onto the imaging surface of the imaging element which is a two-dimensional plane.
- the position X obj of the point on the subject in the three-dimensional space (object surface) can be obtained by performing the perspective projection inverse transformation of Eq. (4) with respect to the position X img_W on the imaging surface of the subject.
- equation (6) the position (variable) X obj of the point on the subject in the three-dimensional space is apparently erased, and the position X img_W on the image plane of the subject at the time of wide-angle imaging as a certain two-dimensional plane. Therefore , it is converted to the position X img_T on the imaging surface of the subject at the time of telescopic imaging as another two-dimensional plane.
- the position X obj on the subject in the three-dimensional space is once determined.
- the process of obtaining a virtual image obtained by virtual imaging based on the information obtained by actual imaging includes actual imaging, generation of a virtual subject (model), and virtual imaging. Consists of.
- a subject in a physical space is subjected to fluoroscopic projection conversion on an imaging element by an optical system (physical lens optical system) such as a physical lens in an imaging device to obtain a two-dimensional image.
- An captured image is generated.
- the perspective projection conversion in actual imaging is performed optically with the physical imaging position (physical imaging position) of the imaging apparatus as a parameter.
- the perspective projection inverse of the equation (4) is calculated by using the distance information from the imaging position to the subject separately obtained by measurement or the like with respect to the captured image obtained by the actual imaging.
- the transformation is performed, and the subject (subject model) in the three-dimensional space is virtually reproduced (generated).
- This virtually reproduced subject is also called a virtual subject (model).
- the virtual subject is (virtually) imaged and a virtual image (virtual image) is generated by performing the perspective projection conversion of the formula (5) on the virtual subject by calculation.
- a virtual imaging position when imaging a virtual subject is specified as a parameter, and the virtual subject is imaged from the virtual imaging position.
- FIG. 12 is a diagram showing an example of an imaging situation when the subject exists on a plurality of object surfaces.
- FIG. 12 as in FIG. 1, the imaging situation is shown by the third trigonometry.
- FIG. 12 shows an imaging situation when the subject exists on a plurality of object surfaces as described above.
- equations (6) and (8) are used to set the position X img_W of the subject on the imaging surface as an actual imaging, for example, in a short-distance imaging, as a virtual imaging. For example, it can be converted to the position X img_T of the subject on the imaging surface in long-distance imaging.
- FIG. 13 is a top view showing a state of wide-angle imaging in which a subject is imaged from an imaging position close to the subject using a wide-angle lens in the imaging situation of FIG.
- FIG. 14 is a top view showing a state of telephoto imaging in which a subject is imaged from an imaging position far from the subject using a telephoto lens in the imaging situation of FIG.
- FIGS. 9 and 10 are views in which an object surface and an imaging surface are added to the second subject with respect to FIGS. 9 and 10.
- the first object surface is the object surface of the first subject
- the second object surface is the object surface of the second subject. Since the second subject is imaged at the same time as the first subject as the background of the first subject, the imaging surface is the same for the first subject and the second subject in FIGS. 13 and 14, respectively.
- the second subject whose position on the second object surface is X obj2 is imaged from the imaging position separated by the object distance L obj_W2.
- the image distance at the time of wide-angle imaging is L img_W
- the second subject whose position on the second object surface is X obj2 is imaged from the imaging position separated by the object distance L obj_T2.
- the image distance at the time of telephoto imaging is L img_T
- the perspective projection inverse transformation equation (9) By applying the perspective projection inverse transformation equation (3) to the wide-angle imaging of FIG. 13, the perspective projection inverse transformation equation (9) can be obtained.
- X obj2 L obj_W2 / L img_W ⁇ X img_W2 ... (9)
- the perspective projection conversion equation (10) By applying the perspective projection conversion equation (2) to the telephoto imaging of FIG. 14, the perspective projection conversion equation (10) can be obtained.
- X img_T2 L img_T / L obj_T2 ⁇ X obj2 ... (10)
- Equation (11) can be obtained by substituting X obj2 on the left side of equation (9) into X obj2 on the right side of equation (10).
- X img_T2 (L img_T / L obj_T2 ) ⁇ (L obj_W2 / L img_W ) ⁇ X img_W2 ... (11)
- equation (11) can be a simple proportional equation of equation (13).
- wide-angle imaging using a wide-angle lens here, telephoto using a telephoto lens from the position X img_W2 on the imaging surface in short-range imaging from a short distance.
- Imaging here the position X img_T2 on the imaging plane in long-distance imaging from a long distance can be obtained.
- the equation (8) is applied to the second object surface.
- the pixels of the captured image obtained by short-distance imaging can be used as virtual imaging, for example, of the virtual image obtained by long-distance imaging. It can be mapped to pixels.
- FIG. 15 is a diagram showing a state of short-distance imaging in the imaging situation of FIG. 12 and an captured image obtained by the short-distance imaging.
- a of FIG. 15 is a top view showing a state of short-distance imaging in which a subject is imaged from an imaging position close to the subject using a wide-angle lens in the imaging situation of FIG.
- FIG. 15B is a plan view showing an captured image obtained by the short-distance imaging of FIG. 15A, which is equivalent to a front view of the imaging surface viewed from the front.
- FIG. 15A is a diagram in which a dotted line as an auxiliary line passing through the center of the lens from each end point of the first subject and the second subject is added to FIG.
- FIG. 16 is a diagram showing a state of long-distance imaging in the imaging situation of FIG. 12 and an captured image obtained by the long-distance imaging.
- a of FIG. 16 is a top view showing a state of long-distance imaging in which a subject is imaged from an imaging position far from the subject by using a telephoto lens in the imaging situation of FIG.
- B of FIG. 16 is a plan view showing an captured image obtained by the long-distance imaging of A of FIG. 16, and is equivalent to a front view of the imaging surface viewed from the front as in the case of FIG.
- FIG. 16A is a diagram in which a dotted line as an auxiliary line passing through the center of the lens from each end point of the first subject and the second subject is added to FIG.
- the size of the second subject on the imaging surface is larger in the long-distance imaging of FIG. 16 than in the short-distance imaging of FIG.
- the phenomenon that the size of the second subject on the imaging surface is larger in the long-distance imaging than in the short-distance imaging is due to the perspective method as described in FIGS. 2 and 3.
- FIG. 17 is a top view showing a state of imaging in which a top view of A in FIG. 15 and a top view of A in FIG. 16 are superposed with a part omitted.
- the portion M of the second subject is imaged in the long-distance imaging, but is not imaged in the shadow of the first subject in the short-distance imaging.
- occlusion that is, a state in which the first subject as the subject on the front side hides the second subject as the subject on the back side and cannot be seen may occur.
- Part M of the second subject is visible in long-distance imaging, but in short-distance imaging, it becomes an occlusion that is hidden behind the first subject and cannot be seen. It is also referred to as an occlusion part (missing part) of the part M of the second subject which is occlusion in this way.
- the portion M of the second subject as the occlusion portion is not imaged. Therefore, when a virtual image obtained by a long-distance imaging as a virtual imaging is generated by using the equations (8) and (13) based on the captured image obtained by the short-distance imaging, the virtual image is used.
- the portion M of the second subject as the occlusion portion is omitted because the pixel value cannot be obtained.
- FIG. 18 is a diagram illustrating mapping of pixel values in the case of generating a virtual image obtained by long-distance imaging as virtual imaging based on an captured image obtained by short-distance imaging as actual imaging. ..
- the portion M of the second subject is behind the first subject and is an occlusion portion.
- the position of the second subject X The pixel value of the pixel of img_W2 is the position of the first subject X img_T and the position of the second subject of the virtual image (long-distance image) using the equations (8) and (13). Each is mapped as the pixel value of the pixel of X img_T2.
- the pixel value of the pixel in which the portion M of the second subject appears should be mapped to the shaded portion.
- the portion M of the second subject is not captured, and the pixel value of the portion M of the second subject cannot be obtained. Therefore, in the virtual image on the lower side of FIG. 18, the pixel value of the portion M of the second subject cannot be mapped to the shaded portion, and the pixel value is missing.
- the pixel value is missing in the occlusion portion such as the portion M of the second subject.
- FIG. 19 is another diagram illustrating mapping of pixel values in the case of generating a virtual image obtained by long-distance imaging as virtual imaging based on an captured image obtained by short-distance imaging as actual imaging. Is.
- the image picW is an captured image obtained by short-range imaging as an actual imaging
- the image picT is a virtual image obtained by long-distance imaging as a virtual imaging.
- the horizontal axis of the two-dimensional coordinates represent the position X Img_W and X Img_W2 lateral captured image PicW
- the vertical axis represents the position X Img_T and X Img_T2 lateral virtual image PICT.
- the straight line L1 represents the equation (8)
- the straight line L2 represents the equation (13).
- Position X Img_W of pixels of the first subjects in the captured image PicW (pixel values), the position X Img_T pixels (pixels of the first object of the virtual image picT obtained by Equation (8) the position X Img_W as input Value).
- Pixel position X Img_W2 the second object of the captured image picW is mapped to the pixel position X Img_T2 the second object of the virtual image picT obtained by the equation (13) the position X Img_W2 as input.
- the shaded part is an occlusion part where the corresponding part is not shown in the captured image picW, and the pixel (pixel value) is missing.
- FIG. 20 is a diagram illustrating an example of a method of complementing the occlusion portion that supplements the pixels of the occlusion portion.
- a method of complementing the occlusion portion for example, there is a method of interpolating the pixel (pixel value) of the occlusion portion by using the pixels in the vicinity of the occlusion portion.
- a method of interpolating pixels for example, an arbitrary method such as a nearest neighbor method, a bilinear method, or a bicubic method can be adopted.
- the pixel values of nearby pixels are used as they are as the pixel values of the pixels in the occlusion portion.
- the average value of the pixel values of the peripheral pixels around the pixel of the occlusion portion is used as the pixel value of the pixel of the occlusion portion.
- the interpolation value obtained by performing three-dimensional interpolation using the pixel values of the peripheral pixels around the pixel of the occlusion portion is used as the pixel value of the pixel of the occlusion portion.
- the occlusion part is, for example, a monotonous wall surface image
- the occlusion part is complemented by interpolation using pixels in the vicinity of the occlusion part, so that the image is taken from the virtual image pickup position where the virtual image is taken. It is possible to generate a virtual image that is (almost) similar to the image obtained when this is done.
- the virtual image is also referred to as a highly reproducible virtual image.
- the periphery of the occlusion portion is used.
- a method of interpolating the occlusion portion can be adopted by duplicating an area of a certain area.
- the method of interpolating the pixels of the occlusion part using the pixels in the vicinity of the occlusion part presupposes that the assumption that the occlusion part will have the same image as the vicinity of the occlusion part is correct.
- the occlusion part does not have the same image as the vicinity of the occlusion part (when the occlusion part is peculiar compared to the vicinity of the occlusion part), the pixels in the vicinity of the occlusion part are used to use the occlusion part. With the method of interpolating the pixels of, it may not be possible to obtain a virtual image with high reproducibility.
- the graffiti part of the wall with some graffiti is the occlusion part
- the graffiti cannot be reproduced by the method of interpolating the pixels of the occlusion part using the pixels in the vicinity of the occlusion part.
- a virtual image with high reproducibility cannot be obtained.
- the occlusion part does not have the same image as the vicinity of the occlusion part, in order to obtain a highly reproducible virtual image, as an actual imaging, in addition to the main imaging (original imaging), the occlusion that occurs in the main imaging
- Auxiliary imaging can be performed from an imaging position different from the imaging position of the main imaging so that the portion is captured. Then, the captured image obtained by the auxiliary imaging can be used to complement the occlusion portion generated in the main imaging.
- FIG. 20 is a top view illustrating a main imaging and an auxiliary imaging performed as actual imaging of the first subject and the second subject.
- the actual imaging with the position p201 as the imaging position is performed as the main imaging, and the actual imaging with the positions p202 and p203 shifted to the left and right from the position p201 as the imaging position is performed as an auxiliary imaging, respectively. It is said.
- the portion M of the second subject which is the occlusion portion
- the portion M of the second subject which is the occlusion portion in the main imaging
- a virtual image obtained by virtual imaging is generated based on the captured image obtained by the main imaging from the imaging position p201, and in the virtual image, the portion M (pixel value of) of the second subject serving as the occlusion portion is generated. ) Is complemented with the captured images obtained by the auxiliary imaging from the imaging positions p202 and p203, so that a virtual image with high reproducibility can be obtained.
- the main imaging and the auxiliary imaging can be performed simultaneously or at different timings using a plurality of imaging devices.
- the main imaging and the auxiliary imaging can be performed by using one imaging device such as a multi-camera having a plurality of imaging systems.
- main imaging and the auxiliary imaging can be performed at different timings by using one imaging device having one imaging system.
- auxiliary imaging can be performed before or after the main imaging.
- Complementation of the occlusion part can be performed using only a part of information such as the color and texture of the captured image obtained by the auxiliary imaging. Furthermore, the complementation of the occlusion part can also be performed in combination with other methods.
- the complementation of the occlusion portion is performed by using the captured image obtained by the auxiliary imaging, and the captured image obtained by another main imaging, for example, the captured image obtained by the main imaging performed in the past. Can be done using.
- the second subject which is the background of the first subject
- a famous building such as Tokyo Tower
- the captured image may be stored in an image library such as a stock photo service.
- the occlusion portion can be performed by using the captured image of the same prominent building, which has been captured in the past and stored in the image library.
- the occlusion portion can be complemented by using an image published on a network such as the Internet, for example, a photograph published on a website that provides a map search service.
- Complementation of the occlusion part can be performed using images as well as data (information) other than images.
- the second subject which is the background of the first subject
- information such as the shape of the building, the surface finishing method, the paint color, and the like can be obtained as architectural data related to the building of the building.
- the occlusion part can be complemented by estimating the pixel value of the occlusion part using such building data.
- the building that is, here, the second subject needs to be identified.
- the second subject can be specified by image recognition of the captured image in which the second subject is captured, by specifying the position where the actual imaging of the captured image is performed, and the like.
- the position where the actual image was taken can be specified by referring to the metadata of the captured image such as EXIF (Exchangeable image file format) information.
- the actual imaging is performed in a situation where the subject is illuminated by a light source such as sunlight.
- the occlusion part is complemented using the past captured image (captured image captured in the past) or architectural data, the light source (illumination by) at the time of actual imaging is not reflected in the occlusion part.
- the color of the occlusion part (the part that was) is changed. , It can be an unnatural color compared to the color of other parts.
- the intensity and color temperature of the light that illuminates the subject are affected by the weather. If meteorological data is available, the weather at the time of actual imaging is identified from the meteorological data, and the light that illuminates the subject at the time of actual imaging performed under sunlight from that weather. Illumination light information such as intensity and color temperature can be estimated.
- the occlusion part is complemented by using the past captured images and architectural data, and the color of the occlusion part is the color when the subject is illuminated by the light represented by the illumination light information. Color tone correction can be performed.
- the color of the occlusion part can be made a natural color compared with the color of other parts, and a virtual image with high reproducibility can be obtained.
- complementation of the occlusion part can be performed using, for example, a learning model in which machine learning has been performed.
- the captured images obtained by actually performing short-distance imaging and long-distance imaging can be used as training data, for example.
- Learning the learning model so that the captured image obtained by the short-distance imaging performed as the actual imaging is input and the image of the occlusion part of the virtual image obtained by the long-distance imaging performed as the virtual imaging is output. It can be carried out.
- the occlusion part can be complemented by the image.
- the complementing method for complementing the occlusion part is not particularly limited. However, by adopting a complementary method that can be performed with one imaging device or even a small number of imaging devices, it is possible to suppress a decrease in mobility at the imaging site and to perform a desired position (virtual imaging position). ), The image (virtual image) captured from) can be easily obtained. In particular, by adopting a complementary method that can be performed by one imaging device, it is possible to maximize the mobility at the imaging site.
- FIG. 21 is a diagram illustrating another example of the process of obtaining a virtual image obtained by virtual imaging based on the information obtained by actual imaging.
- the process of obtaining a virtual image obtained by virtual imaging based on the information obtained by actual imaging is the same as in the case of FIG. 11, the actual imaging, the generation of the virtual subject (model), and the process of obtaining the virtual subject (model). , Consists of virtual imaging. However, in FIG. 21, in the case of FIG. 11, the complement of the occlusion portion is added.
- an captured image (actual captured image), which is a two-dimensional image, is generated (captured) as in the case of FIG.
- the virtual subject as a corrected model is reproduced (generated) from the captured image obtained by the actual imaging by using the distance information from the imaging position of the actual imaging to the subject and the coping model information. Will be done.
- the coping model information is knowledge information for coping with occlusion, for example, a captured image captured in the past (past captured image), an captured image obtained by auxiliary imaging (auxiliary captured image), and architectural data. , And one or more such as meteorological data.
- the virtual subject is generated by performing the perspective projection inverse transformation of the captured image obtained in the actual imaging using the distance information.
- the virtual imaging position is given as a parameter, and in the virtual imaging performed later, the imaging portion of the virtual subject imaged from the virtual imaging position is specified.
- the missing portion where the pixel (pixel value) of the captured image is missing that is, the occlusion portion that is occlusion when the virtual subject is viewed from the virtual imaging position.
- a virtual image is generated by the perspective projection conversion as in the case of FIG.
- the target of the perspective projection conversion is not the virtual subject itself generated by performing the perspective projection inverse conversion of the captured image obtained in the actual imaging, but the occlusion of the virtual subject. It is different from the case of FIG. 11 in that it is a virtual subject as a corrected model in which a part is complemented.
- the corrected model is (virtually) imaged from the virtual imaging position by performing perspective projection conversion on the corrected model, and a virtual image (virtual image) is generated.
- a virtual image virtual image
- the range of complementation can be minimized by performing the complementation only for the occlusion part that is occlusion when the virtual subject is viewed from the virtual imaging position.
- the auxiliary imaging can be performed in the minimum necessary range, and deterioration of mobility during imaging can be suppressed.
- auxiliary imaging by imaging a range slightly wider than the minimum necessary, it is possible to finely correct the virtual imaging position after imaging.
- FIG. 22 is a plan view showing an example of a captured image and a virtual image.
- imaging is performed when one of the short-distance imaging and the long-distance imaging is actually captured and the other is imaged.
- the generation method for generating the virtual image to be generated has been described. That is, a position that is moved along the optical axis of the imaging device at the time of imaging and differs only in the imaging distance from the imaging position where the captured image is actually captured is set as the virtual imaging position, and the virtual imaging position is based on the captured image.
- the generation method for generating a virtual image captured from is described.
- the above-mentioned method for generating a virtual image can also be applied when a position moved from the imaging position of the captured image in a direction not along the optical axis of the imaging device is set as the virtual imaging position. That is, the above-mentioned virtual image generation is performed along the optical axis of the imaging device, in addition to the case of generating a virtual image of the subject from the imaging position of the captured image and the position moved along the optical axis of the imaging device. It can also be applied when a virtual image (another virtual image) of an image of a subject is generated from a position moved in a non-direction.
- the optical axis of the image pickup device When the optical axis of the image pickup device is directed toward the subject and the subject is imaged, the position moved along the optical axis of the image pickup device at the time of actual imaging from the image pickup position of the captured image is actually taken as the virtual image pickup position.
- the optical axis of the image pickup apparatus coincides between the image pickup and the virtual image pickup.
- the imaging device when the position moved from the imaging position of the captured image in a direction not along the optical axis of the imaging device at the time of actual imaging is set as the virtual imaging position, the imaging device is used in the actual imaging and the virtual imaging.
- the optical axis of is different.
- the virtual imaging position is a position moved from the imaging position of the captured image in a direction not along the optical axis of the imaging device, for example, actual imaging is performed in the imaging situation of FIG. 1, and virtual imaging is performed. Is applicable in the case of performing the above in the imaging situation shown in FIG.
- a of FIG. 22 shows an captured image obtained by actual imaging in the imaging situation of FIG. 1.
- FIG. 22 shows a bird's-eye view image as an captured image obtained by actual imaging in the imaging situation of FIG.
- C in FIG. 22 performs virtual imaging on a virtual subject generated by using distance information based on the image captured in A in FIG. 22 with the imaging position in the imaging state of FIG. 4 as the virtual imaging position.
- the virtual image obtained (generated) by the above is shown.
- the occlusion part that is occlusion when the virtual subject is viewed from the virtual image pickup position is complemented by using the coping model information, and the corrected virtual subject after the complementation is complemented.
- a method of complementing the occlusion part a method of interpolating the occlusion part using pixels in the vicinity of the occlusion part described above, a method of using an captured image obtained by auxiliary imaging, or a method of using an image captured in the past can be used. It is possible to adopt a method to be used, a method to use a learning model learned by machine learning, a method to use architectural data, and the like.
- FIG. 23 is a diagram illustrating a method of expressing a virtual imaging position when performing virtual imaging.
- the actual imaging position for imaging is determined by physically (actually) installing the imaging device.
- a virtual imaging position is required in addition to the imaging position of the actual imaging, and it is necessary to specify the virtual imaging position.
- a designation method for designating the virtual imaging position for example, a method of automatically designating a position moved by a predetermined distance in a predetermined direction with respect to the imaging position as a virtual imaging position can be adopted. Further, as a designation method for designating the virtual imaging position, for example, a method of having the user specify the position can be adopted.
- the UI when the user specifies the virtual imaging position will be described below, but before that, the expression method of the virtual imaging position will be described.
- the virtual imaging position is represented by a spherical coordinate system (coordinates) centered (origin) at the position of the subject (which intersects the optical axis of the imaging device). do.
- the optical axis of the imaging device (physically existing physical imaging device), that is, the intersection of the optical axis of the imaging device (optical system) and the subject is defined as the center of the subject.
- the optical axis of the image pickup device passes through the center of the image pickup device of the image pickup device and coincides with a straight line perpendicular to the image pickup device.
- optical axis connecting the center of the image sensor of the image sensor and the center of the subject is called the physical optical axis
- the optical axis connecting the virtual imaging position and the center of the subject is called the virtual optical axis. ..
- the virtual imaging position is the amount of rotation in the azimuth direction (azimuth angle) ⁇ v and the amount of rotation in the elevation angle direction (elevation angle) with respect to the physical optical axis of the virtual optical axis. ) ⁇ v and the distance r v between the subject on the virtual optical axis and the virtual imaging position can be expressed.
- the distance r r represents the distance between the subject on the physical optical axis and the imaging position.
- the distance r v shown in the top view represents a distance along a virtual optical axis, not a distance component on a plane.
- FIG. 24 is a plan view showing an example of a UI that is operated when the user specifies a virtual imaging position.
- the UI has operation buttons such as a C button, a TOP button, a BTM button, a LEFT button, a RIGHT button, a SHORT button, a LONG button, a TELE button, and a WIDE button.
- buttons such as a C button, a TOP button, a BTM button, a LEFT button, a RIGHT button, a SHORT button, a LONG button, a TELE button, and a WIDE button.
- the UI can be configured by using a rotary dial, a joystick, a touch panel, and other operation units in addition to the operation buttons. Further, when the UI is configured by using the operation buttons, the arrangement of the operation buttons is not limited to the arrangement shown in FIG. 24.
- An imaging device to which this technology is applied can generate a virtual image similar to the captured image obtained by imaging a subject from a virtual imaging position in real time and output it to a display unit such as a viewfinder in real time.
- the display unit can display the virtual image as a so-called through image in real time.
- the azimuth angle ⁇ v can be specified by operating the LEFT button or the RIGHT button.
- the azimuth angle ⁇ v changes in the negative direction by a predetermined fixed amount.
- the azimuth angle ⁇ v changes in the positive direction by a predetermined fixed amount.
- the elevation angle ⁇ v can be specified by operating the TOP button or the BTM button.
- the TOP button When the TOP button is pressed, the elevation angle ⁇ v changes in the positive direction by a predetermined fixed amount.
- the BTM button When the BTM button is pressed, the elevation angle ⁇ v changes in the negative direction by a predetermined fixed amount.
- the distance r v can be specified by operating the SHORT button or the LONG button.
- the distance r v changes in the negative direction by a predetermined fixed amount or a fixed magnification.
- the LONG button is pressed, the distance r v changes in the positive direction by a predetermined fixed amount or a fixed magnification.
- the UI includes the C button, TOP button, BTM button, LEFT button, RIGHT button, SHORT button, LONG button, and other virtual imaging positions related to the designation of the virtual imaging position.
- a TELE button and a WIDE button for designating the focal length of the virtual imaging device (hereinafter, also referred to as the virtual focal length) when performing the virtual imaging of the above are provided.
- the virtual focal length changes in the direction of increasing by a predetermined fixed amount or a fixed magnification. Further, when the WIDE button is pressed, the virtual focal length changes in the direction of shortening by a predetermined fixed amount or a fixed magnification.
- the image distances L img_W and L img_T of the equations (4) and (5) are determined according to the virtual focal length.
- the method of changing the azimuth angle ⁇ v, etc. with respect to the operation of the operation buttons of the UI is not limited to the above.
- the virtual imaging position and virtual focal length such as the azimuth angle ⁇ v are continuously changed, and the long press is continued. Convenience can be enhanced by changing the virtual imaging position such as the azimuth angle ⁇ v and the amount of change in the virtual focal length according to the time.
- the method of specifying the virtual imaging position is not limited to the method of operating the UI.
- the position of the gazing point is designated (set) as the virtual imaging position.
- the imaging device when displaying a virtual image obtained by virtual imaging from a virtual imaging position specified by a UI operation or the like in real time, the occlusion portion is displayed so that the user can recognize it. can do.
- the accuracy of the information of the complement portion complemented by the occlusion portion may be inferior to the image obtained by actually capturing the image from the virtual imaging position.
- the virtual image is displayed so that the user can recognize the occlusion portion that becomes the occlusion in the virtual image obtained by the virtual imaging from the virtual imaging position. , Can be displayed on the display unit.
- the user of the imaging device can recognize which part of the subject is the occlusion part by looking at the virtual image displayed on the display unit. Then, by recognizing which part of the subject becomes the occlusion part, the user of the imaging device can consider the actual imaging position of the image so that the important part of the subject does not become the occlusion part for the user. can. That is, the imaging position can be considered so that an important part of the subject is captured by the user in the captured image obtained by the actual imaging.
- the part of the subject that becomes the occlusion part becomes.
- Auxiliary imaging can be performed so that it can be captured.
- the occlusion part in the virtual image is displayed in a specific color.
- a method, a method of reversing the gradation of the occlusion portion at a predetermined cycle such as 1 second, or the like can be adopted.
- FIG. 25 is a block diagram showing a configuration example of an embodiment of an imaging device such as a digital camera to which the present technology is applied.
- the image pickup device 100 includes an image pickup optical system 2, an image pickup element 3, a distance sensor 5, an inverse conversion unit 7, a correction unit 9, a conversion unit 11, a display unit 13, UI 15, a storage unit 17, and recording units 21 to 23. And, it has an output unit 24.
- the imaging device 100 can be applied to imaging both moving images and still images.
- the image pickup optical system 2 collects the light from the subject on the image pickup element 3 to form an image. As a result, the subject in the three-dimensional space is transformed into a perspective projection on the image sensor 3.
- the image sensor 3 receives the light from the image pickup optical system 2 and performs photoelectric conversion to generate an image pickup image 4 which is a two-dimensional image having a pixel value corresponding to the amount of light received, and the inverse conversion unit 7 Supply to.
- the distance sensor 5 measures and outputs distance information 6 to each point of the subject.
- the distance information 6 output by the distance sensor 5 is supplied to the inverse conversion unit 7.
- the distance information 6 of the subject can be measured by an external device and supplied to the inverse transformation unit 7.
- the image pickup apparatus 100 can be configured without providing the distance sensor 5.
- the inverse transformation unit 7 performs perspective projection inverse transformation of the captured image 4 from the image sensor 3 using the distance information 6 from the distance sensor 5, and generates and outputs a virtual subject as three-dimensional data 8.
- the correction unit 9 complements the occlusion portion of the virtual subject as the three-dimensional data 8 output by the inverse transformation unit 7, and outputs the complemented virtual subject as the corrected model 10.
- the conversion unit 11 performs perspective projection conversion of the corrected model 10 output by the correction unit 9, and outputs a virtual image 12 which is a two-dimensional image obtained as a result.
- the display unit 13 displays the virtual image 12 output by the conversion unit 11.
- the display unit 13 can display the virtual image 12 in real time.
- the UI 15 is configured as shown in FIG. 24, for example, and is operated by a user who is an imager, for example, an image pickup device 100. The user can operate the UI 15 for designating the virtual imaging position 16 while viewing the virtual image displayed on the display unit 13.
- the UI 15 sets and outputs the virtual imaging position 16 according to the user's operation.
- the correction unit 9 complements the occlusion portion that becomes an occlusion when the virtual subject is viewed from the virtual imaging position 16 output by the UI 15.
- the correction unit 9 specifies an occlusion portion that becomes an occlusion when the virtual subject is viewed from the virtual imaging position 16. After that, the correction unit 9 complements the occlusion portion, and the complemented virtual subject is output as the corrected model 10.
- the virtual image 12 which is a two-dimensional image obtained by imaging the corrected model 10 output by the correction unit 9 from the virtual imaging position 16 output by the UI 15 is subjected to the perspective projection conversion of the corrected model 10. Will be generated.
- the display unit 13 displays the virtual image 12 obtained by imaging the corrected model 10 from the virtual imaging position 16 set according to the operation of the UI 15 by the user in real time.
- the user can specify the virtual imaging position 16 from which the desired virtual image 12 can be obtained by operating the UI 15 while looking at the virtual image 12 displayed on the display unit 13.
- the occlusion portion can be complemented by interpolating the occlusion portion using pixels in the vicinity of the occlusion portion. Further, the complementation of the occlusion portion can be performed by obtaining the past captured image 18, the building data 19, the weather data 20, etc. as the coping model information from the outside and using the coping model information.
- the occlusion part can be complemented by using an captured image obtained by auxiliary imaging, a machine-learned learning model, or the like as other coping model information.
- the inverse transformation unit 7 displays the three-dimensional data 8 generated from the captured image 4 obtained by the auxiliary imaging.
- the virtual subject is stored in the storage unit 17.
- the storage unit 17 stores the virtual subject as the three-dimensional data 8 generated from the captured image 4 obtained by the auxiliary imaging in the inverse transformation unit 7.
- the virtual subject as the three-dimensional data 8 stored in the storage unit 17 and generated from the captured image 4 obtained by the auxiliary imaging is obtained by the main imaging performed after the auxiliary imaging in the correction unit 9. It can be used to complement the occlusion portion that is occlusion when the virtual subject as the three-dimensional data 8 generated from the captured image 4 is viewed from the virtual imaging position 16.
- the inverse transformation unit 7 performs a virtual subject as three-dimensional data 8 generated from the captured image 4 obtained by the main imaging. Is recorded in the recording unit 23.
- the recording unit 23 stores the virtual subject as the three-dimensional data 8 generated from the captured image 4 obtained by the main imaging in the inverse transformation unit 7.
- Complementing the occlusion portion of the virtual subject as the three-dimensional data 8 generated from the captured image 4 obtained by the main imaging recorded in the recording unit 23 is supplemented in the correction unit 9 after the main imaging. This can be performed using a virtual subject as three-dimensional data 8 generated from the captured image 4 obtained by imaging.
- the auxiliary imaging when the auxiliary imaging is performed after the main imaging, the captured image obtained by the main imaging is waited for the auxiliary imaging to be performed after the main imaging.
- the occlusion part of the virtual subject as the three-dimensional data 8 generated from is complemented.
- the occlusion portion When the occlusion portion is complemented, it is difficult to generate the virtual image 12 from the captured image 4 obtained by the main imaging in real time. Therefore, when the generation of a virtual image in real time is required, the auxiliary imaging needs to be performed prior to the main imaging, not after the main imaging.
- the recording unit 21 records the virtual image 12 output by the conversion unit 11.
- the virtual image 12 recorded in the recording unit 21 can be output to the display unit 13 and the output unit 24.
- the recording unit 22 records the corrected model 10 output by the correction unit 9.
- a wide range portion (virtual imaging position 16) that is somewhat wider than the occlusion portion, including an occlusion portion that is occlusion when the virtual subject is viewed from the virtual imaging position 16 from the UI 15. It changes slightly, and when the virtual subject is viewed from the virtual imaging position 16 after the change, the portion including the portion of the virtual subject that becomes a new occlusion portion) can be complemented.
- the recording unit 22 can record the corrected model 10 in which such a wide range of complement is performed.
- the corrected model 10 recorded in the recording unit 22 with the complement of a wide range portion is used as the target of the perspective projection conversion of the conversion unit 11, and the virtual imaging position 16 is finely modified (finely adjusted). 12 can be generated. Therefore, after capturing the captured image 4 that is the basis of the corrected model 10 in which the wide area is complemented, the corrected model 10 in which the wide area is complemented recorded in the recording unit 22 is used. It is possible to generate a virtual image 12 in which the virtual imaging position 16 is slightly modified.
- the recording unit 23 records the virtual subject as the three-dimensional data 8 output by the inverse transformation unit 7, that is, the virtual subject before the occlusion portion is complemented by the correction unit 9.
- the virtual subject recorded in the recording unit 23 is processed to confirm the authenticity of, for example, when the virtual image 12 is used in news or the like and the authenticity of a part of the virtual image 12 is questioned. No, so to speak, it can be referred to as true data.
- the recording unit 23 contains the captured image that is the source of the generation of the virtual subject as the three-dimensional data 8 together with the virtual subject as the three-dimensional data 8 or instead of the virtual subject as the three-dimensional data 8. 4 can be recorded.
- the output unit 24 is an I / F (interface) that outputs data to the outside of the image pickup apparatus 100, and outputs the virtual image 12 output by the conversion unit 11 to the outside in real time.
- the output unit 24 delivers the virtual image 12 to the external device in real time. be able to.
- the virtual image 12 is output from the output unit 24 to the external display unit.
- the virtual image 12 is displayed in real time on the external display unit.
- the inverse transformation unit 7 performs three-dimensional transformation of the captured image 4 from the image sensor 3 by using the distance information 6 from the distance sensor 5.
- a virtual subject as data 8 is generated.
- the correction unit 9 occludes when the virtual subject as the three-dimensional data 8 generated by the inverse transformation unit 7 is viewed from the virtual imaging position 16 from the UI 15 using the coping model information such as the past captured image 18.
- the complemented virtual subject as the corrected model 10 in which the virtual subject is corrected is obtained.
- the conversion unit 11 uses the corrected model 10 obtained by the correction unit 9 to generate a virtual image 12 obtained by capturing the corrected model 10 from the virtual imaging position 16 by perspective projection conversion.
- the inverse transformation unit 7, the correction unit 9, and the conversion unit 11 use the distance information 6 from the imaging position to the subject and the coping model information to obtain the imaging position from the captured image 4 in which the subject is captured from the imaging position. It can be said that the generation unit that generates the virtual image 12 in which the subject is imaged from the virtual imaging position 16 different from the above is configured.
- FIG. 26 is a flowchart illustrating an example of processing in the generation unit.
- step S1 the generation unit uses the distance information 6 and the coping model information (knowledge information) to deal with the occlusion of the past captured image 18, etc., from the captured image 4 to a virtual imaging position different from the captured image 4.
- a virtual image 12 captured from 16 is generated.
- step S11 the inverse transformation unit 7 of the generation unit generates the virtual subject as the three-dimensional data 8 by performing the perspective projection inverse transformation of the captured image 4 using the distance information 6.
- step S12 the inverse transformation unit 7 of the generation unit generates the virtual subject as the three-dimensional data 8 by performing the perspective projection inverse transformation of the captured image 4 using the distance information 6. The process proceeds to step S12.
- step S12 when the correction unit 9 sees the virtual subject as the three-dimensional data 8 generated by the inverse transformation unit 7 from the virtual imaging position 16 using the coping model information such as the past captured image 18.
- a corrected model 10 three-dimensional data 8 with the occlusion portion complemented
- step S13 the conversion unit 11 uses the corrected model 10 generated by the correction unit 9 to generate a virtual image obtained by capturing the corrected model 10 from the virtual imaging position 16 by perspective projection conversion.
- the imaging device 100 for example, even in a situation where it is difficult to image a subject from a desired imaging position (viewpoint), an image captured from a certain imaging position (viewpoint) that can be imaged and an image captured.
- the distance information from the imaging position to the subject and the coping model information as auxiliary information other than the distance information obtained separately, from the virtual imaging position as a desired imaging position different from the actual imaging position. It is possible to generate a virtual image captured in a pseudo manner. Therefore, an image (virtual image) captured from a desired position can be easily obtained.
- a virtual image as if the image was taken from a position behind the wall surface in front of the person is captured. Can be generated.
- the image pickup device 100 in an image pickup situation in which the user of the image pickup device 100 cannot approach the subject, such as an image pickup situation in which the user of the image pickup device 100 takes an image of the outside through a window while riding in a room or a vehicle, the user It is possible to generate a virtual image as if it was taken close to the subject.
- a virtual image such as the bird's-eye view image shown in FIG. 5 can be generated without using a stepladder, a drone, or the like.
- the image pickup apparatus 100 for example, when the subject is a person and the line of sight of the person is not facing the image pickup device, the position at the tip of the line of sight is set as the virtual image pickup position, so that the virtual image of the camera's line of sight is virtual. Images can be generated.
- the imaging device 100 by setting the virtual imaging position to the position of the eyeball of the head of the user who is the imager, it is possible to generate a virtual image in which the state seen from the user's viewpoint is captured. By displaying such a virtual image on a spectacle-type display, electronic spectacles without parallax can be constructed.
- a series of processes of the inverse conversion unit 7, the correction unit 9, and the conversion unit 11 that constitute the above-mentioned generation unit can be performed by hardware or software.
- the programs constituting the software are installed on a general-purpose computer or the like.
- FIG. 27 is a block diagram showing a configuration example of an embodiment of a computer in which a program for executing the above-mentioned series of processes is installed.
- the program can be recorded in advance on the hard disk 905 or ROM 903 as a recording medium built in the computer.
- the program can be stored (recorded) in the removable recording medium 911 driven by the drive 909.
- a removable recording medium 911 can be provided as so-called package software.
- examples of the removable recording medium 911 include a flexible disc, a CD-ROM (Compact Disc Read Only Memory), a MO (Magneto Optical) disc, a DVD (Digital Versatile Disc), a magnetic disc, and a semiconductor memory.
- the program can be installed on the computer from the removable recording medium 911 as described above, or can be downloaded to the computer via a communication network or a broadcasting network and installed on the built-in hard disk 905. That is, for example, the program transfers wirelessly from a download site to a computer via an artificial satellite for digital satellite broadcasting, or transfers to a computer by wire via a network such as LAN (Local Area Network) or the Internet. be able to.
- LAN Local Area Network
- the computer has a built-in CPU (Central Processing Unit) 902, and the input / output interface 910 is connected to the CPU 902 via the bus 901.
- CPU Central Processing Unit
- the CPU 902 executes a program stored in the ROM (Read Only Memory) 903 accordingly. .. Alternatively, the CPU 902 loads the program stored in the hard disk 905 into the RAM (Random Access Memory) 904 and executes it.
- ROM Read Only Memory
- the CPU 902 performs processing according to the above-mentioned flowchart or processing performed according to the above-mentioned block diagram configuration. Then, the CPU 902 outputs the processing result from the output unit 906, transmits it from the communication unit 908, and further records it on the hard disk 905, if necessary, via the input / output interface 910.
- the input unit 907 is composed of a keyboard, a mouse, a microphone, and the like. Further, the output unit 906 is composed of an LCD (Liquid Crystal Display), a speaker, or the like.
- LCD Liquid Crystal Display
- the processing performed by the computer according to the program does not necessarily have to be performed in chronological order in the order described as the flowchart. That is, the processing performed by the computer according to the program also includes processing executed in parallel or individually (for example, parallel processing or processing by an object).
- the program may be processed by one computer (processor) or may be distributed processed by a plurality of computers. Further, the program may be transferred to a distant computer and executed.
- the system means a set of a plurality of components (devices, modules (parts), etc.), and it does not matter whether all the components are in the same housing. Therefore, a plurality of devices housed in separate housings and connected via a network, and a device in which a plurality of modules are housed in one housing are both systems. ..
- this technology can have a cloud computing configuration in which one function is shared by a plurality of devices via a network and processed jointly.
- each step described in the above flowchart can be executed by one device or shared by a plurality of devices.
- one step includes a plurality of processes
- the plurality of processes included in the one step can be executed by one device or shared by a plurality of devices.
- a generation unit that uses distance information from the imaging position to the subject and model information to generate a virtual image of the subject from a virtual imaging position different from the imaging position from the captured image obtained by imaging the subject from the imaging position.
- the generator A virtual subject is generated by performing a perspective projection inverse transformation of the captured image using the distance information.
- a corrected model in which the virtual subject is corrected is generated.
- ⁇ 7> The imaging device according to ⁇ 1>, further comprising a UI (User Interface) for designating the virtual imaging position.
- UI User Interface
- ⁇ 8> The imaging device according to any one of ⁇ 1> to ⁇ 7>, which outputs the virtual image to a display unit in real time.
- the UI is A first operation unit operated when determining the center of the spherical coordinate system representing the virtual imaging position, and A second operation unit operated when changing the azimuth angle of the virtual imaging position in the spherical coordinate system, A third operation unit operated when changing the elevation angle of the virtual imaging position in the spherical coordinate system,
- the imaging device according to ⁇ 7> which has a fourth operating unit that is operated when changing the distance between the center of the spherical coordinate system and the virtual imaging position.
- the UI further includes a fifth operation unit that is operated when changing the focal length of the virtual imaging device when performing virtual imaging from the virtual imaging position. ..
- ⁇ 11> The imaging device according to ⁇ 10>, wherein the UI continuously changes the virtual imaging position or the focal length while the operation of any of the first to fifth operation units is continued.
- ⁇ 12> The imaging device according to ⁇ 10>, wherein the UI changes the amount of change in the virtual imaging position or the focal length according to the time during which the operation of any of the first to fifth operation units is continued. .. ⁇ 13>
- the UI is an imaging device according to any one of ⁇ 1> to ⁇ 12>, which designates a gazing point at which a user is gazing at the virtual imaging position.
- Imaging method including. ⁇ 15> A generation unit that uses distance information from the imaging position to the subject and model information to generate a virtual image of the subject from a virtual imaging position different from the imaging position from the captured image obtained by imaging the subject from the imaging position.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Computer Graphics (AREA)
- General Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Computing Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Studio Devices (AREA)
- User Interface Of Digital Computer (AREA)
- Image Processing (AREA)
Abstract
Description
・・・(1)
・・・(2)
・・・(3)
・・・(4)
・・・(5)
・・・(6)
・・・(7)
・・・(8)
・・・(9)
・・・(10)
・・・(11)
・・・(12)
・・・(13)
撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する生成部を備える
撮像装置。
<2>
前記生成部は、前記距離情報と前記モデル情報とを用いて、前記撮像画像から補正済みモデルを生成し、前記補正済みモデルを用いて、前記仮想画像を生成する
<1>に記載の撮像装置。
<3>
前記モデル情報は、オクルージョンに対処するための知識情報である
<1>又は<2>に記載の撮像装置。
<4>
前記生成部は、
前記距離情報を用いて、前記撮像画像の透視投影逆変換を行うことにより、仮想被写体を生成し、
前記モデル情報を用いて、前記仮想撮像位置から前記仮想被写体を見たときにオクルージョンになっているオクルージョン部分を補完することで、前記仮想被写体を補正した補正済みモデルを生成し、
前記補正済みモデルを用いて、透視投影変換により、前記補正済みモデルを前記仮想撮像位置から撮像した前記仮想画像を生成する
<3>に記載の撮像装置。
<5>
前記仮想被写体又は前記補正済みモデルを記録する記録部をさらに備える
<4>に記載の撮像装置。
<6>
前記モデル情報は、過去に撮像された前記撮像画像、建築に関する建築データ、及び、気象に関する気象データの1つ以上を含む
<3>ないし<5>のいずれかに記載の撮像装置。
<7>
前記仮想撮像位置を指定するUI(User Interface)をさらに備える
<1>に記載の撮像装置。
<8>
前記仮想画像を、表示部にリアルタイムで出力する
<1>ないし<7>のいずれかに記載の撮像装置。
<9>
前記UIは、
前記仮想撮像位置を表現する球面座標系の中心を決定するときに操作される第1の操作部と、
前記球面座標系における前記仮想撮像位置の方位角を変化させるときに操作される第2の操作部と、
前記球面座標系における前記仮想撮像位置の仰角を変化させるときに操作される第3の操作部と、
前記球面座標系の中心と前記仮想撮像位置との間の距離を変化させるときに操作される第4の操作部と
を有する
<7>に記載の撮像装置。
<10>
前記UIは、前記仮想撮像位置からの仮想的な撮像を行うときの仮想的な撮像装置の焦点距離を変化させるときに操作される第5の操作部をさらに有する
<9>に記載の撮像装置。
<11>
前記UIは、前記第1ないし第5の操作部のいずれかの操作が継続されている間、前記仮想撮像位置又は前記焦点距離を連続的に変化させる
<10>に記載の撮像装置。
<12>
前記UIは、前記第1ないし第5の操作部のいずれかの操作が継続されている時間に応じて、前記仮想撮像位置又は前記焦点距離の変化量を変化させる
<10>に記載の撮像装置。
<13>
前記UIは、ユーザが注視している注視点を、前記仮想撮像位置に指定する
<1>ないし<12>のいずれかに記載の撮像装置。
<14>
撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する
ことを含む撮像方法。
<15>
撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する生成部
としてコンピュータを機能させるためのプログラム。
Claims (15)
- 撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する生成部を備える
撮像装置。 - 前記生成部は、前記距離情報と前記モデル情報とを用いて、前記撮像画像から補正済みモデルを生成し、前記補正済みモデルを用いて、前記仮想画像を生成する
請求項1に記載の撮像装置。 - 前記モデル情報は、オクルージョンに対処するための知識情報である
請求項1に記載の撮像装置。 - 前記生成部は、
前記距離情報を用いて、前記撮像画像の透視投影逆変換を行うことにより、仮想被写体を生成し、
前記モデル情報を用いて、前記仮想撮像位置から前記仮想被写体を見たときにオクルージョンになっているオクルージョン部分を補完することで、前記仮想被写体を補正した補正済みモデルを生成し、
前記補正済みモデルを用いて、透視投影変換により、前記補正済みモデルを前記仮想撮像位置から撮像した前記仮想画像を生成する
請求項3に記載の撮像装置。 - 前記仮想被写体又は前記補正済みモデルを記録する記録部をさらに備える
請求項4に記載の撮像装置。 - 前記モデル情報は、過去に撮像された前記撮像画像、建築に関する建築データ、及び、気象に関する気象データの1つ以上を含む
請求項3に記載の撮像装置。 - 前記仮想撮像位置を指定するUI(User Interface)をさらに備える
請求項1に記載の撮像装置。 - 前記仮想画像を、表示部にリアルタイムで出力する
請求項1に記載の撮像装置。 - 前記UIは、
前記仮想撮像位置を表現する球面座標系の中心を決定するときに操作される第1の操作部と、
前記球面座標系における前記仮想撮像位置の方位角を変化させるときに操作される第2の操作部と、
前記球面座標系における前記仮想撮像位置の仰角を変化させるときに操作される第3の操作部と、
前記球面座標系の中心と前記仮想撮像位置との間の距離を変化させるときに操作される第4の操作部と
を有する
請求項7に記載の撮像装置。 - 前記UIは、前記仮想撮像位置からの仮想的な撮像を行うときの仮想的な撮像装置の焦点距離を変化させるときに操作される第5の操作部をさらに有する
請求項9に記載の撮像装置。 - 前記UIは、前記第1ないし第5の操作部のいずれかの操作が継続されている間、前記仮想撮像位置又は前記焦点距離を連続的に変化させる
請求項10に記載の撮像装置。 - 前記UIは、前記第1ないし第5の操作部のいずれかの操作が継続されている時間に応じて、前記仮想撮像位置又は前記焦点距離の変化量を変化させる
請求項10に記載の撮像装置。 - 前記UIは、ユーザが注視している注視点を、前記仮想撮像位置に指定する
請求項1に記載の撮像装置。 - 撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する
ことを含む撮像方法。 - 撮像位置から被写体までの距離情報とモデル情報とを用いて、前記撮像位置から前記被写体を撮像した撮像画像から、前記撮像位置と異なる仮想撮像位置から前記被写体を撮像した仮想画像を生成する生成部
としてコンピュータを機能させるためのプログラム。
Priority Applications (5)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US17/782,851 US20230005213A1 (en) | 2020-01-23 | 2021-01-08 | Imaging apparatus, imaging method, and program |
EP21744897.6A EP4075789A4 (en) | 2020-01-23 | 2021-01-08 | IMAGING DEVICE, IMAGING PROCEDURE AND PROGRAM |
CN202180009485.9A CN115004683A (zh) | 2020-01-23 | 2021-01-08 | 成像装置、成像方法和程序 |
JP2021573062A JPWO2021149509A1 (ja) | 2020-01-23 | 2021-01-08 | |
JP2024092068A JP2024114712A (ja) | 2020-01-23 | 2024-06-06 | 撮像装置、撮像方法、及び、プログラム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2020-008936 | 2020-01-23 | ||
JP2020008936 | 2020-01-23 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021149509A1 true WO2021149509A1 (ja) | 2021-07-29 |
Family
ID=76992931
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/000445 WO2021149509A1 (ja) | 2020-01-23 | 2021-01-08 | 撮像装置、撮像方法、及び、プログラム |
Country Status (5)
Country | Link |
---|---|
US (1) | US20230005213A1 (ja) |
EP (1) | EP4075789A4 (ja) |
JP (2) | JPWO2021149509A1 (ja) |
CN (1) | CN115004683A (ja) |
WO (1) | WO2021149509A1 (ja) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114167891A (zh) * | 2021-11-29 | 2022-03-11 | 湖南汽车工程职业学院 | 一种基于无人机的地面数据采集处理系统 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005100367A (ja) * | 2003-09-02 | 2005-04-14 | Fuji Photo Film Co Ltd | 画像生成装置、画像生成方法、及び画像生成プログラム |
JP2010177921A (ja) * | 2009-01-28 | 2010-08-12 | Fujifilm Corp | 立体撮像装置および立体撮像方法 |
WO2011096136A1 (ja) * | 2010-02-02 | 2011-08-11 | コニカミノルタホールディングス株式会社 | 疑似画像生成装置および疑似画像生成方法 |
JP2019103126A (ja) | 2018-10-04 | 2019-06-24 | キヤノン株式会社 | カメラシステム、カメラ制御装置、カメラ制御方法及びプログラム |
Family Cites Families (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2004114224A1 (ja) * | 2003-06-20 | 2004-12-29 | Nippon Telegraph And Telephone Corporation | 仮想視点画像生成方法及び3次元画像表示方法並びに装置 |
WO2011096252A1 (ja) * | 2010-02-02 | 2011-08-11 | コニカミノルタホールディングス株式会社 | 画像処理システム、画像処理方法、およびプログラム |
US8675090B2 (en) * | 2010-12-15 | 2014-03-18 | Panasonic Corporation | Image generating apparatus, image generating method, and recording medium |
JP6112824B2 (ja) * | 2012-02-28 | 2017-04-12 | キヤノン株式会社 | 画像処理方法および装置、プログラム。 |
JP6029380B2 (ja) * | 2012-08-14 | 2016-11-24 | キヤノン株式会社 | 画像処理装置、画像処理装置を具備する撮像装置、画像処理方法、及びプログラム |
JP6214236B2 (ja) * | 2013-03-05 | 2017-10-18 | キヤノン株式会社 | 画像処理装置、撮像装置、画像処理方法、及びプログラム |
JP6395423B2 (ja) * | 2014-04-04 | 2018-09-26 | キヤノン株式会社 | 画像処理装置、制御方法及びプログラム |
JP6543062B2 (ja) * | 2015-03-23 | 2019-07-10 | キヤノン株式会社 | 画像処理装置、画像処理方法 |
US20170094227A1 (en) * | 2015-09-25 | 2017-03-30 | Northrop Grumman Systems Corporation | Three-dimensional spatial-awareness vision system |
US10970915B2 (en) * | 2017-01-06 | 2021-04-06 | Canon Kabushiki Kaisha | Virtual viewpoint setting apparatus that sets a virtual viewpoint according to a determined common image capturing area of a plurality of image capturing apparatuses, and related setting method and storage medium |
JP6415675B1 (ja) * | 2017-12-04 | 2018-10-31 | キヤノン株式会社 | 生成装置、生成方法及びプログラム |
TWI657409B (zh) * | 2017-12-27 | 2019-04-21 | 財團法人工業技術研究院 | 虛擬導引圖示與真實影像之疊合裝置及其相關疊合方法 |
KR102622714B1 (ko) * | 2018-04-08 | 2024-01-08 | 디티에스, 인코포레이티드 | 앰비소닉 깊이 추출 |
EP3651119A1 (en) * | 2018-11-06 | 2020-05-13 | Koninklijke Philips N.V. | Disparity estimation from a wide angle image |
US10957027B2 (en) * | 2019-03-26 | 2021-03-23 | Intel Corporation | Virtual view interpolation between camera views for immersive visual experience |
US11644890B2 (en) * | 2021-02-11 | 2023-05-09 | Qualcomm Incorporated | Image capturing in extended reality environments |
-
2021
- 2021-01-08 CN CN202180009485.9A patent/CN115004683A/zh active Pending
- 2021-01-08 EP EP21744897.6A patent/EP4075789A4/en active Pending
- 2021-01-08 JP JP2021573062A patent/JPWO2021149509A1/ja active Pending
- 2021-01-08 WO PCT/JP2021/000445 patent/WO2021149509A1/ja active Application Filing
- 2021-01-08 US US17/782,851 patent/US20230005213A1/en active Pending
-
2024
- 2024-06-06 JP JP2024092068A patent/JP2024114712A/ja active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2005100367A (ja) * | 2003-09-02 | 2005-04-14 | Fuji Photo Film Co Ltd | 画像生成装置、画像生成方法、及び画像生成プログラム |
JP2010177921A (ja) * | 2009-01-28 | 2010-08-12 | Fujifilm Corp | 立体撮像装置および立体撮像方法 |
WO2011096136A1 (ja) * | 2010-02-02 | 2011-08-11 | コニカミノルタホールディングス株式会社 | 疑似画像生成装置および疑似画像生成方法 |
JP2019103126A (ja) | 2018-10-04 | 2019-06-24 | キヤノン株式会社 | カメラシステム、カメラ制御装置、カメラ制御方法及びプログラム |
Non-Patent Citations (1)
Title |
---|
See also references of EP4075789A4 |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114167891A (zh) * | 2021-11-29 | 2022-03-11 | 湖南汽车工程职业学院 | 一种基于无人机的地面数据采集处理系统 |
CN114167891B (zh) * | 2021-11-29 | 2022-08-16 | 湖南汽车工程职业学院 | 一种基于无人机的地面数据采集处理系统 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2021149509A1 (ja) | 2021-07-29 |
CN115004683A (zh) | 2022-09-02 |
US20230005213A1 (en) | 2023-01-05 |
JP2024114712A (ja) | 2024-08-23 |
EP4075789A4 (en) | 2023-05-03 |
EP4075789A1 (en) | 2022-10-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109887003B (zh) | 一种用于进行三维跟踪初始化的方法与设备 | |
EP2328125B1 (en) | Image splicing method and device | |
US10547822B2 (en) | Image processing apparatus and method to generate high-definition viewpoint interpolation image | |
US20130335535A1 (en) | Digital 3d camera using periodic illumination | |
US20110025830A1 (en) | Methods, systems, and computer-readable storage media for generating stereoscopic content via depth map creation | |
US9813693B1 (en) | Accounting for perspective effects in images | |
US10545215B2 (en) | 4D camera tracking and optical stabilization | |
JP2008140271A (ja) | 対話装置及びその方法 | |
JP2010109783A (ja) | 電子カメラ | |
CN110751735B (zh) | 一种基于增强现实的远程指导的方法与设备 | |
CN111737518A (zh) | 基于三维场景模型的图像展示方法、装置及电子设备 | |
JP2024114712A (ja) | 撮像装置、撮像方法、及び、プログラム | |
US20220329770A1 (en) | Information processing apparatus, video generation method and program | |
WO2011014421A2 (en) | Methods, systems, and computer-readable storage media for generating stereoscopic content via depth map creation | |
CN113643414A (zh) | 一种三维图像生成方法、装置、电子设备及存储介质 | |
CN113724303A (zh) | 点云与图像匹配方法、装置、电子设备和存储介质 | |
US20240087157A1 (en) | Image processing method, recording medium, image processing apparatus, and image processing system | |
US10154241B2 (en) | Depth map based perspective correction in digital photos | |
JP2023546739A (ja) | シーンの3次元モデルを生成するための方法、装置、およびシステム | |
US20240348928A1 (en) | Image display method, device and electronic device for panorama shooting to improve the user's visual experience | |
CN111402404A (zh) | 全景图补全方法、装置、计算机可读存储介质及电子设备 | |
CN109978945B (zh) | 一种增强现实的信息处理方法和装置 | |
CN110286906A (zh) | 用户界面显示方法、装置、存储介质与移动终端 | |
US20240054667A1 (en) | High dynamic range viewpoint synthesis | |
CN116450002A (zh) | Vr图像处理方法、装置、电子设备及可读存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21744897 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2021573062 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2021744897 Country of ref document: EP |
|
ENP | Entry into the national phase |
Ref document number: 2021744897 Country of ref document: EP Effective date: 20220714 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |